def __init__(self, vcf_file, fast_forward=0) :
self.indexOf = broad.COLUMN_MAP
globes.printColumnWarning( vcf_file, self.indexOf )
self.fin = open( vcf_file, "rb" )
self.patients = broad.getPatients( self.fin )
self.allow_absent = False
self.group_repeats = False
self.iterator = self.iterate(fast_forward)
def parseForINDEL( indel_file ) :
print indel_file
fin = open( indel_file )
(path,ext) = indel_file.split('.',1)
fname = path.split('/')[-1]
fout = open( "%s/intermediate_data/sift/input/%s_sift_input.csv" \
% (globes.DATA_DIR, fname), 'wb' )
print fout
#fast-forward through header lines
patients = broad.getPatients( fin )
indexOf = broad.COLUMN_MAP
for dataline in fin :
splt = dataline.strip().split('\t')
col_keys = ['chrom','pos','mut','ref']
chrom,pos,mut,ref = [ splt[ indexOf[k] ] for k in col_keys ]
dinfo = broad.makeInfoDict( splt[ indexOf["info"] ] )
try :
strand = dinfo["refseq.transcriptStrand"]
except KeyError :
try :
strand = dinfo["refseq.transcriptStrand_1"]
except KeyError : #dont have it guess '+'
strand = '+'
##raise Exception("what the f**k: %s" % splt[ indexOf["info"] ] )
if strand == "+" : strand = 1
elif strand == "-" : strand = -1
else : raise Exception("Strand is not + or - ??")
isInsertion = len(ref) == 1 and len(mut) > 1
isDeletion = len(ref) > 1 and len(mut) == 1
if isInsertion :
start = int(pos)
end = start
allele = mut
elif isDeletion :
start = int(pos)
end = start + (len(ref)-len(mut))
allele = '-/'
else : assert False
fout.write( "%s,%d,%d,%d,%s\n" % (chrom,start,end,strand,allele) )
fout.close()
fin.close()
def separateOutputToFamilies() :
fin = open( "%s/seattle/input/indel_input.vcf" % (globes.INT_DIR) )
patients = broad.getPatients( fin )
fouts = [ open("%s/indels_by_fam/%s.tsv" % (globes.OUT_DIR, \
pat.replace('/','-')), 'wb' ) \
for pat in patients ]
fin.close()
#errrgg so I can re-get out the original read data
finin = open( "%s/seattle/input/indel_input.vcf" % (globes.INT_DIR) )
patients = broad.getPatients( finin )
finin_splt = finin.readline().strip().split('\t')
fin = open( "%s/seattle/output/indel_output.tsv" % (globes.INT_DIR) )
column_splt = fin.readline().strip().split('\t')
bp = indexOf["sampleAlleles"]
column_splt = column_splt[:bp] + ["originalBroadCall"] + column_splt[bp:]
new_columns = "\t".join( column_splt )
for fout in fouts :
fout.write( "%s\n" % new_columns )
for line in fin :
#ignore the comment lines at the end
if '#' in line : continue
#get the necessary column values
splt = line.strip().split('\t')
cols = ["chromosome","position","refBase","sampleGenotype"]
chrom,pos,refBase,sampleGTs = [ splt[ indexOf[c] ] for c in cols ]
sampleGTs = sampleGTs.split(',')
#find line in input file that corresponds to the output line
num_incs = 0
while True :
cols = ["chrom","pos"]
values = [ finin_splt[ broad.COLUMN_MAP[c] ] for c in cols ]
finin_chrom, finin_pos = values
finin_calls = finin_splt[ broad.COLUMN_MAP["calls"]: ]
if pos == finin_pos and chrom == finin_chrom :
break
else :
num_incs += 1
finin_splt = finin.readline().strip().split('\t')
#The output may have multiple lines for each input line, corresponding
#to the different transcripts. This means that if 'line' no longer
#matches the finin_line, we should only have to jump next once
assert num_incs <= 1
#isMutated is a function that takes a GT from the output file
#and determines if it is a mutation
isInsertion = '-' in refBase
if isInsertion :
l,r = refBase.split('-')
isMutated = lambda gt : \
not gt == '%s/%s' % (l,l) and not gt == "N/N"
else :
isMutated = lambda gt : \
not refBase in gt.split('/')[1] and not gt == 'N/N'
num_mutations = 0
for i,(fout,gt) in enumerate( zip(fouts,sampleGTs) ) :
if isMutated(gt) :
num_mutations += 1
#splt_copy = list(splt)
#print splt_copy
splt[ indexOf["sampleGenotype"] ] = "%s" % (gt)
newline = "\t".join( splt[:bp] + [ finin_calls[i] ] + splt[bp:] )
fout.write( "%s\n" % newline )
#because of indel.indelUniqueToDisease
assert num_mutations == 1
[f.close() for f in fouts]
fin.close()
